Research On Some Key Issues Of MES In Complex Information Environment

As materialization center and manufacturing information center, shop floor must satisfy the requirements for agility to adapt to modern manufacturing mode. This paper carries out a thorough study on some key technologies in Manufacturing Execution System (MES) which helps optimize manufacturing processes in practical complex information fulfilled manufacturing environments. Theories and methodologies are put forward systematically to solve MES reconfiguration, scheduling, task execution monitor and exception handling issues. Complex information environment is the practical running environment for MES. The origination of complex information is analyzed, and the mutual effects between MES and the complex information environment are explored. Some strategies and technologies are recommended for complex information processing involved in MES, which may offer some guidelines to MES research works. In order to meet the demands of structural agility and run-time agility, a distributed MES framework based on Agent technology is put forward in view of common features of advanced shop floor control, which satisfies reconfiguration, reusability, and scalabe requirements of manufacturing systems. Dynamic reconfiguration is responsively fast, and is creative in organization for shop floor management in a MES based on agile manufacturing cells. A new dynamic cell reconfiguration model based on incomplete knowledge and evolutionary game theory is introduced in this paper. By adopting both the top-down and down-top approaches, stable optimized reconfigurable manufacturing cells could be produced oriented to uncertain demand objectives. A simple universal coevolutionary algorithm based on pointer-based integer coding schema is implemented to find the optimal solution of the reconfiguration model. Meanwhile, applicability and efficacy of the recommended method are verified by a case study. MES scheduling within complex information circumstances needs appropriate scheduling mechanisms and methods. A MES dynamic scheduling model based on time decomposition is set up according to the balance policy between scheduling and execution which makes full use of real-time information and reasonably simplifies the non-deterministic scheduling problems. Furthermore, a general genetic algorithm-based dynamic scheduling framework for MES that can be reconfigured and extended is presented. Under such a framework, a hybrid genetic algorithm (HGA) for MES scheduling in complex information environment is proposed, in which some key elements of HGA such as code and decode schema, cross and mute operators are thoroughly investigated. Also an experimental dynamic MES scheduling prototype system is developed, and related algorithms are verified by numeric simulation. In addition, a new HGA with immune mechanism is proposed for further improvement. As a link between practical shop floor execution and MES reconfigurations, scheduling and even top level planning systems, task execution monitor and exception handling play an important role in MES. After establishing an interactive model for task execution monitor with respect to other components in MES, key issues in MES monitor are presented by combined analysis of static scheduling data and dynamic execution perturbation. Moreover, a multi-agent-based MES task execution monitor method is proposed, and an example execution is carried out. The fast exception handling method is based on intelligent behaviors of agents including information seeking, binary subjective information evaluation and irregular task reallocation, whose efficacy is proven by an illustration. In the end, the conclusion is drawn, and research trends on MES are anticipated.